Title: Relational Constraints in Pharmaceutical Experiments : Impact on the Design of the Experiments

Speaker: Prof. Manisha Pal
Department of Statistics, Faculty of Science
St. Xavier’s University Kolkata, India

Date: Wed, 19 June 2024

Time: 13:00-14:00 PM

Room: M306

Abstract:

Design of experiments has vast application in different fields, like engineering, pharmaceutical, biomedical, environmental and epidemiological research. In some of these areas, especially in pharmaceutical research, the response is defined, not in terms of the actual amounts of different constituents used, but in terms of their relative proportions. Such a response function defines a mixture model. Theoretically, the proportions of components in a mixture can vary between 0 and 1, subject to their sum being equal to 1. However, in many practical situations, this is not realistic. For example, in preparing a fruit punch there generally is a restriction on the proportions of different fruit-juices used. There are also situations where the components of a mixture can be classified into a number of classes, depending on their type and function. There may be constraints on the proportions of these classes used in the mixture. Such constraints are called relational constraints.
In this presentation, I shall talk about relational constraints in pharmaceutical experiments, and how these impact the design of the experiments.